|Publication number||US7339822 B2|
|Application number||US 10/313,738|
|Publication date||Mar 4, 2008|
|Filing date||Dec 6, 2002|
|Priority date||Dec 6, 2002|
|Also published as||CN1720585A, CN100552808C, DE60327257D1, EP1586096A2, EP1586096B1, US7319630, US20040109354, US20050101236, WO2004053883A2, WO2004053883A3|
|Publication number||10313738, 313738, US 7339822 B2, US 7339822B2, US-B2-7339822, US7339822 B2, US7339822B2|
|Inventors||Chi-Ming Wang, Kuo-Lung Chen, Shouchang Tsao|
|Original Assignee||Sandisk Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (40), Non-Patent Citations (10), Referenced by (4), Classifications (7), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The present invention relates to programming of nonvolatile erasable programmable memories and more specifically, a technique to more rapidly program these memory cells by applying a bias voltage to source nodes of the memory cells.
Memory and storage is one of the key technology areas that is enabling the growth in the information age. With the rapid growth in the Internet, World Wide Web (WWW), wireless phones, personal digital assistants (PDAs), digital cameras, digital camcorders, digital music players, computers, networks, and more, there is continually a need for better memory and storage technology.
A particular type of memory is the nonvolatile memory. A nonvolatile memory retains its memory or stored state even when power is removed. Some types of nonvolatile erasable programmable memories include Flash, EEPROM, EPROM, MRAM, FRAM, ferroelectric, and magnetic memories. Some nonvolatile storage products include Flash disk drives, CompactFlash™ (CF) cards, MultiMedia cards (MMC), secure digital (SD) cards, Flash PC cards (e.g., ATA Flash cards), SmartMedia cards, personal tags (P-Tag), and memory sticks.
A widely used type of semiconductor memory storage cell is the Flash memory cell. Some types of floating gate memory cells include Flash, EEPROM, and EPROM. There are other types of memory cell technologies such as those mentioned above. Floating gate memory cells such as Flash are discussed as merely an example. The discussion in this application would also apply to other memory technologies other than floating gate technology with the appropriate modifications.
Memory cells are configured or programmed to a desired configured state. In particular, electric charge is placed on or removed from the floating gate of a Flash memory cell to put the cell into two or more stored states. One state is a programmed state and another state is an erased state. A Flash memory cell can be used to represent at least two binary states, a 0 or a 1. A Flash memory cell can also store more than two binary states, such as a 00, 01, 10, or 11. This cell can store multiple states and may be referred to as a multistate memory cell, a multilevel, or multibit memory cell. This allows the manufacture of higher density memories without increasing the number of memory cells since each memory cell can represent more than a single bit. The cell may have more than one programmed state. For example, for a memory cell capable of representing two bits, there will be three programmed states and an erased state.
Despite the success of nonvolatile memories, there also continues to be a need to improve the technology. It is desirable to improve the density, performance, speed, durability, and reliability of these memories. It is also desirable to reduce power consumption and reduce the cost per bit of storage. One aspect of nonvolatile memories is the circuitry and techniques used to write or program data into the cells. Latch circuits may be used for this purpose. Conventional data latches are directly connected to power lines, VDD and VSS. However connecting these power supplies to latches may cause longer delay and more power consumptions. It is desirable that the circuitry provides the proper voltages to the cells to support the need operations, be compact in size, have low power consumption, and be relatively easy to adjust as needed.
As can be appreciated, there is a need for improving the circuitry and techniques for operating on memory cells.
The invention is a current-limited latch circuit, which may be used within a nonvolatile memory integrated circuit for decoding, programming, erase, and other operations. In a specific implementation, there are a number of latches connected together in parallel between two power supply lines. The latch uses a current mirroring scheme to limit current supplied to the latch. This reduces a difference of the two supplies, positive voltage, ground, or negative voltages, during data changes. The circuit provides smaller device sizes and fast speeds when data changes in the latch, while also providing lower power consumption. The technique provides greater benefits as the voltage difference between the two power supplies is greater.
In one specific implementation, the invention is an integrated circuit including an array of nonvolatile memory cells arranged in rows and columns and a plurality of latch circuits. The memory cells may include Flash, EEPROM, EPROM, floating gate, HE, or DFGSSI cells. The latch circuits are connected to the array of memory cells. The latch circuits are connected in parallel between a upper supply line and a lower supply line. Each latch circuit includes a upper pull-up circuit block connected to the upper supply line and a lower pull-down circuit block connected to the lower supply line. The supply lines are supplied by current-limited supplies, which may be generated on-chip. Further, in an embodiment, in a first mode, the upper supply line is placed at a positive voltage. In a second mode, the upper supply line is placed at a negative voltage and the lower supply line is also placed at a negative voltage, more negative than the upper supply line.
The upper pull-up circuit block includes a first p-channel transistor having a source connected to the upper supply line, a gate connected to a first node, and a source connected to a second node. The upper pull-up circuit block further includes a second p-channel transistor having a source connected to the upper supply line, a gate connected to the second node, and a source connected to the first node.
The lower pull-down circuit block includes a first n-channel transistor having a source connected to the lower supply line, a gate connected to a first node, and a source connected to a second node. The lower pull-down circuit block further includes a second n-channel transistor having a source connected to the lower supply line, a gate connected to the second node, and a source connected to the first node.
The latch includes a first n-channel transistor connected between a first and second node, and has a gate connected to a bias voltage line. The latch has a second n-channel transistor connected between a third and fourth node, and a gate connected to the bias voltage line. The first and second n-channel transistors may be deep n-well devices, each formed in a p-type diffusion tub that is within an n-well tub.
The latch may further include a first p-channel transistor connected between the second node and a fifth node, and a second p-channel transistor connected between the fourth node and a sixth node. During normal operation, the fifth node and sixth node provide complementary logic signals (in any voltage level). The p-channel transistors may be triple n-well devices, especially if negative supplies are used.
In embodiment, the upper supply line of the latch circuit is connected to a positive voltage source of 10 volts or greater and the lower supply line is connected to ground. During operation, a voltage difference between voltages at the upper supply line and lower supply lines may be about 10 volts or greater.
A voltage supplied to the upper supply line is generated by a first on-chip pump circuit, and a voltage supplied to the lower supply line is generated by a second on-chip pump circuit. The current from the upper supply line may be limited to about 10 microamps and a current to the lower supply line may be limited to about 10 microamps. The amount the current is limited by may be any level, and depends on the specific application or need. In other embodiments, the current limiter may limit to 100 microamps or less.
In a further embodiment, the invention is an integrated circuit having nonvolatile memory cells and a number of latches, one latch for each row of memory cells. The latch is laid out on a substrate of the integrated circuit in the same pitch as a row of memory cells. The latch may store positive and negative voltages.
Other objects, features, and advantages of the present invention will become apparent upon consideration of the following detailed description and the accompanying drawings, in which like reference designations represent like features throughout the figures.
This electronic system architecture includes a processor or microprocessor 21 connected to a system bus 23, along with random access, main system memory 25, and at least one or more input-output devices 27, such as a keyboard, monitor, modem, and the like. Another main computer system component that is connected to a typical computer system bus 23 is an amount of long-term, nonvolatile memory 29. In contrast to volatile memory such as DRAM (dynamic RAM) or SRAM (static RAM), nonvolatile memory retains its stored state even after power is removed from the device. Typically, such a memory is a disk drive using magnetic or optical technology with a capacity of megabytes, gigabytes, or terabytes of data storage. This data is retrieved into the system volatile memory 25 for use in current processing, and can be easily supplemented, changed, or altered.
One aspect of the invention is the substitution of a specific type of semiconductor memory system for the disk drive but without having to sacrifice nonvolatility, ease of erasing and rewriting data into the memory, speed of access, low cost and reliability. This is accomplished by employing an one or more electrically erasable programmable read only memories (e.g., Flash or EEPROMs) integrated circuits. Integrated circuits are sometimes called chips. This type of memory has additional advantages of requiring less power to operate, and of being lighter in weight than a hard-disk drive magnetic media memory, thus being especially suited for battery-operated portable computers. Such nonvolatile semiconductor memories include Flash disk drives, CompactFlash (™) cards, SmartMedia (™) cards, personal tags (P-Tag), multimedia cards, secure digital (SD) cards, and memory sticks (R).
The bulk storage memory 29 is constructed of a memory controller 31, connected to the computer system bus 23, and an array 33 of Flash or EEPROM integrated circuit chips. Data and instructions are communicated from the controller 31 to the Flash or EEPROM array 33 primarily over a data line 35. Similarly, data and status signals are communicated from the Flash or EEPROM 33 to the controller 31 over data lines 37. Data lines 35 and 37 may be serial or parallel, depending on the implementation. Other control and status circuits between the controller 31 and the EEPROM array 33 are not shown in
The nonvolatile memory integrated circuits may also be combined with other integrated circuits or components such as controllers, microprocessors, random access memories (RAM), or I/O devices, to form a nonvolatile memory system. The controller and memory may be on separate integrated circuits or a memory integrated circuit may incorporate the controller. The memory may reside on multiple, separate integrated circuits. For example, multiple memory integrated circuits may be combined to obtain larger memory sizes.
Further discussion of Flash EEPROM systems and nonvolatile cells and storage is discussed in U.S. Pat. No. 5,602,987, U.S. Pat. No. 5,095,344, U.S. Pat. No. 5,270,979, U.S. Pat. No. 5,380,672, U.S. Pat. 5,712,180, U.S. Pat. No. 5,991,517, U.S. Pat. No. 6,222,762, and U.S. Pat. No. 6,230,233, which are incorporated by reference along with all other references cited in this application.
A memory integrated circuit of a nonvolatile memory system will include a number of memory cells, each of which holds at least one bit of data. Multistate memory cells may also be used, which will allow storage of multiple bits of data in each cell. For example, each memory cell may store two, three, four, five, six, seven, eight, or more bits of data per cell. Memory cells capable of storing multiple bits of data may also be referred to as multilevel cells.
Some types of nonvolatile storage or memory cells are Flash, EEPROM, and EPROM, which are all floating gate type memory cells. The invention may also be applied to other types of memories such as phase-change cells, magnetic cells (MRAM), ferroelectric cells (FRAM), magnetic ferroelectric, and many others.
The memory cells are typically arranged in an array in rows and columns. There may be multiple arrays per integrated circuit. Individual cells are accessed by row and column. Two different organizations for the memory cells are NOR and NAND configurations. The invention is applicable to these configurations as well as other configurations of memory cells.
For example, if the word “source” is reserved for that electrode that is at a lower potential than the drain, then during read operations the line that is connected to the drain of the select transistor is the drain line, and the line that is connected to the source of the memory cell transistor is the source line. The situation is reversed for programming, where the higher voltage is applied to the memory cell side to achieve source side injection.
For an array of NOR memory cells, a number of NOR cells will be connected to the drain line (or source line). This will typically be referred to as a column of the array. Each cell of the column will have a separate word line or row line, which will typically referred to as a row of the array.
In an implementation, both read and memory transistors are n-channel or NMOS type transistors. However, the devices may be other types of transistors including p-channel or PMOS type transistors and others. The read device 211 may be a different device type from the memory device 215. In a specific implementation, the memory device is a floating gate device such as a Flash, EEPROM, or EPROM transistor. However, the memory device may be another type of device such as a phase-change, NRAM, FRAM, magnetic ferroelectric, FeRAM, NROM, MNOS, SONOS, or other device.
In an implementation, the source select transistor, drain select transistor, and memory transistors are n-channel or NMOS type transistors. However, the devices may be other types of transistors including p-channel or PMOS type transistors and others. The drain select device and source select device may be a different device type from the memory device 311, and different from each other. In a specific implementation, the memory device is a floating gate device such as a Flash, EEPROM, or EPROM transistor. However, the memory device may be another type of device such as a phase-change, NRAM, FRAM, magnetic ferroelectric, FeRAM, NROM, MNOS, SONOS, or other device.
In brief, a nonvolatile memory cell is one that retains its stored state, even when power is removed. Some examples of floating gate type memory cells include Flash, EEPROM (also known as E2 or E-squared), and EPROM. A Flash and EEPROM cell is electrically erasable and electrically programmable. An EPROM cell is electrically programmable, and erasable using ultraviolet (UV) light. A floating gate device is programmed or erased by subjecting appropriate nodes to high voltages. These high voltages cause electrons to be added to or removed from the floating gate, which will adjust the threshold voltage or VT of the floating gate device. Some physical mechanisms for causing electrons to move to or from floating gate are hot-electron injection, Fowler-Nordheim tunneling, or band-to-band tunneling.
The high voltage used to program the device is sometimes referred to as the VPP voltage, and the high voltage used to erase the device is sometimes referred to as the VEE voltage. The VPP voltage will vary depending on the process technology and the particular implementation. In a particular implementation, VPP ranges from about 6.5 volts to about 20 volts. In some specific implementations, VPP ranges from 12 volts to 18 volts. The VEE voltage will vary depending on the process technology and the particular implementation. In a particular implementation, VEE ranges from about −12 volts to about −25 volts. In some specific implementations, VEE ranges from −12 volts to −15 volts. In some implementation, the programming voltages may be generated by on-chip circuitry, such as charge pumps or other voltage generator circuits, and in other implementations, the programming voltages may be supplied from a voltage source, external to the integrated circuit.
The floating gate nonvolatile memory device may store a single bit (0 or 1) or multiple bits (e.g., two bits: 00, 01, 10, and 11, or three bits: 000, 001, 010, 011, 100, 101, 110, and 111, or four bits: 0000, 0001, 0010, 0011, 0100, 0101, 0110, 0111, 1000, 1001, 1010, 1011, 1100, 1101, 1110, and 1111). U.S. Pat. No. 5,991,517 discusses further some aspects of single bit and multibit cells. In short, the memory cell will have an erased state and one or more programmed states.
An erased state is when the VT of the device is such that it is off for voltages from about ground to VCC. In other words, erase refers to configuring the floating gate device to have a VT (threshold voltage) of, for example, 0 volts or below. When erased, the floating gate transistor does conduct current even when 1 volt is placed on its gate (i.e., control gate). All the floating gate cells of an integrated circuit may be initialized to an erased state. Furthermore, in an embodiment, the memory cell may need to be erased before it can be programmed.
Erasing occurs by removing electrons from the floating gate of the floating gate transistor. One technique to do this is to ground the control gate and to place the VEE voltage at the drain or the source, or both. The VEE voltage attracts the negatively charged electrons from the floating gate, because they are attracted to the positive voltage. Another technique is to apply a voltage at the drain or source, or both. A more negative voltage (may be referred to as a negative VEE voltage) is applied at the gate. For example, in one embodiment, 2 volts is applied to the source or drain and −19 volts is applied to the control gate.
Generally, erasing occurs by the electrons tunneling out of the floating gate. The time it takes to erase the device depends on various factors including the voltage magnitude difference between the control gate and the source or drain. Generally, the greater the voltage difference, the faster the device will become erased because the electrons will be more strongly attracted to the more positive voltage. However, it is desirable that the VEE voltage is selected so that it does not damage the cell and also allows erase of the selected cell, without disturbing the stored states of adjacent and other memory cells (when in an array).
When storing only one bit, the floating gate device will have in addition to the erased state, only one programmed state. For purposes of this application, a programmed state for a single bit cell is typically when the VT of the device is higher than a designated positive value.
For a multistate cell, the VT is set at specific voltage level indicating it is in a particular state. In other words, depending on what the programmed VT state is, this will indicate a particular stored binary value. For an example of a two-bit memory cell, a VT of 1 volt, plus or minus 0.25 volts may indicate a 01 binary state. A VT of 2 volts, plus or minus 0.25 volts, may indicate a 11 binary state. And a VT of 3 volts, plus or minus 0.25 volts, may indicate a 10 binary state. In this example, Gray coding is used so that with state changes, only one bit changes at a time. In other implementations, other coding techniques may be used.
Programming occurs by adding electrons to the floating gate of the floating gate transistor. Generally, one technique to do this is to place VPP at the control gate and to place ground at the drain or the source, or both. The VPP voltage attracts the negatively charged electrons into the floating gate, because they are attracted to the positive voltage. Using this technique, the electrons tunnel into the floating gate.
Generally, another technique is to place a VPGG voltage (which is a high positive) at control gate and cause a current to flow across the floating gate device's channel region (such as placing 6 volts at the drain and grounding the source). Then hot electrons from the current flow will be attracted and become embedded into the floating gate.
More specifically, to program memory cells, there are two mechanisms, tunneling and hot electron injection. For multistate programming, each write operations includes a sequence of program pulses, each of which is followed by a verify operation.
Channel hot electron injection has both hot electron production and hot electron injection. To produce hot electrons a large lateral field in required. This is provided by a high drain to source voltage. To inject the hot electrons onto the floating gate, a large vertical electric field is used. This electric field is provided by the control gate voltage which in turn couples some of its voltage to the floating gate. In drain side injection, the high vertical field needed for hot electron injection has the side effect of reducing the high lateral field needed for hot electron production. Source side injection does not suffer from the same dilemma, and therefore is more efficient. In both source side and drain side injection, a scattering mechanism is required to divert the momentum of electrons traveling laterally through the channel, so that a lucky few will be scattered vertically towards the floating gate. The normal hot electron impingement that is an innovative feature of this cell, may increase the programming efficiency beyond the improved efficiency of source side injection because the majority of hot electrons will have a momentum that will aid them in surmounting the silicon and silicon oxide energy barrier. It will no longer be necessary for the hot electrons to scatter into a direction that would aid them in surmounting the silicon and silicon oxide energy barrier.
Another mechanism for programming would be Fowler-Nordheim tunneling, as is used in NAND technology. But to use tunneling typically entails giving up the potential benefits of the previously discussed mechanism. Tunneling is generally very slow in comparison to hot electron injection. In the case of tunneling, performance has to be maintained by programming larger numbers of cells in parallel at a cost of larger numbers of peripheral programming blocks.
During each programming pulse the drain voltage is maintained at a constant value in a range from about 3 volts to 6 volts. The control gate voltage for the first programming pulse will have some positive starting value that need to be characterized, and may be adaptively determined even in the field. The transfer gate voltage is a constant envisioned to be in a range from about 6 volts to 10 volts. The select gate or word line voltage is envisioned to be in a range from about 3 volts to 10 volts. The select transistor's threshold voltage would preferably be high enough to shut off any current paths through unselected cells, which prevents disturbing the states of the unselected cells. This is because the optimum select gate voltage for the most efficient source side injection is less than a volt higher that the select gate threshold voltage. For the DFGSSI cell (described below), the drain for programming will be the bit line just adjacent the cell, and the source for programming will be the other bit line residing just to the other side of the drain. It should be noted that the roles of the two bit lines reverses for the read or verify operation where the adjacent bit line to the floating gate is the source. In this naming convention source is that electrode with a lower voltage as compared to the drain. The source voltage for programming may be adaptively controlled by a current limiter such that the instantaneous programming current does not exceed some designated value.
The movement of charge into and out of the floating gate is determined by an electric field magnitude across the tunneling dielectric (which is the gate oxide between the floating gate and channel region): generally the higher the voltage difference between the control gate or the floating gate and the source, the higher the charge transfer into the floating gate. The time it takes to program the device depends on various factors including the voltage magnitude difference between the VPP voltage on control gate. One should note during the programming, the control gate voltage is not necessarily precisely at VPP. VPP is a constant voltage output of a particular charge pump.
Generally, the greater the electric field, the faster the device will become programmed because the electrons will be more strongly attracted to the VPP voltage. However, it is desirable that the maximum programming control gate voltage and the maximum programming drain voltage are selected so that they do not damage the cell and also allow programming of the selected cell, without disturbing the stored states of adjacent and other memory cells (when in an array) on the same control line, bit line, or word line. Further, when the cell is a multistate cell, it may be desirable that the VPP voltage is selected so it allows sufficiently fine resolution in programming the device to the desired VT. For example, the VPP voltage may be applied in pulses so that the devices don't become programmed to too much (i.e., programmed to a VT level above their intended VT level).
The storage element has a select gate line 609, a right control gate line 611, and a left control gate line 613. The right control gate line is connected to a gate or control electrode of a right floating gate transistor (TFGR) 615, and the left control gate line is connected to a gate of a left floating gate transistor (TFGL) 617. The select gate line (word line) is connected a gate of a select transistor (TSEL) 619.
For each storage element 605, there are two floating gate transistors or cells 615 and 617 to store binary data. Each of these floating gate transistors may store a single bit or multiple bits of data. When storing multiple bits of data, each floating gate cell may also be referred to as a multistate, multilevel, or multibit cell, because the cell may be programmed to have more than two VT (threshold voltage) levels. For example, each floating gate transistor may store two bits per cell, four bits per cell, or an even greater number of bits per cell.
Floating gate transistors are selectively configured by placing appropriate voltages on bit lines BL1 and BL2, control gate lines 613 and 611, and select lines 609. The drains and sources of the transistors are connected to bit lines BL1 and BL2, which may be selectively connected to ground, or any another constant voltage or constant current source through transistors 628 and 632. At edges of each control gate segment, every pair of control gate lines can be merged into one electrode in order to fit their control line segment select transistors in the pitch of two floating gates as opposed to the pitch of one floating gate cell as shown in
The array may be organized and subdivided into sectors of cells. This will allow erase of entire sectors of cells at a time, rather than the entire array at once. For example, a sector may contain 512 user bytes. A four-bit-per-cell implementation will have at least 1024 cells per sector. More cells are typically required for error checking and correction (ECC), tracking cells, rotation state, and other features.
In operation, data is read from a selected cell by using the word lines, bit lines, and steering (control) lines when we read TFGR cell 615. Line 609 which is connected to the select gate is biased at 5 volts. CG line 611 is biased at a sense voltage. Line 613 is biased at overdrive level 7 volts to make sure cell 617 is on. And BL1 is at about 1 volt or about 1.5 volts. DL2 is at 0 volts or VSS. Then whether cell 615 conducts or not determines the voltage level on BL1.
In a specific embodiment, the current-limited latch of the invention will be described with respect to the specific storage element structure shown in
The latch of the invention is applicable to integrated circuits where mixed voltage levels are used on-chip, not just for memory integrated circuits. The invention may be used in any logic integrated circuit or mixed signal integrated circuit having both analog and digital circuit portions, just to name some examples.
The circuitry includes a latch portion 702, which has an upper supply pull-up circuit 706 and a lower supply pull-down circuit 709. This application provides some specific sizes of the devices, specific voltage levels, and other specific parameters for a specific implementation of the invention. However, it is understood that other sizes, voltages, and parameter values may be used in other implementations of the invention. For example, the sizes of the transistors may be adjusted to obtain a desired or appropriate result or to operate properly in a particular process.
The latch has two NMOS transistors 704 and 705, where gates of both devices are connected to a VB2 voltage line. In an embodiment, VB2 may be 3 volts of −5 volts, switching between the two as needed. Further in this specific embodiment, both transistors are the same size, having a width/length (W/L) of 32.2/4, where the values are given in microns. Transistors 704 and 705 are included in the embodiment in the figure, but may be removed or optional in other embodiments.
Transistors 704 and 705 are primarily used to protect the other devices, transistors in the lower supply pull-down circuit, from high voltages seen at ER and ERb. Transistors 704 and 705 will help prevent punch through of transistors N13 and N14. Specifically, transistors N13 and N14 of the lower supply pull-down circuit have a breakdown voltage (sometimes referred to as BVDSS). In a specific embodiment, BVDSS for an n-channel device is about 14 volts and for a p-channel device is about 24 volts. Therefore, it is important transistors N13 and N14 are not subjected to the BVDSS voltage or a voltage close to BVDSS. The voltage at ER and ERb is placed across two n-channel devices in series so the drain-to-source for transistors N13 and N14 is reduced, ensuring the voltage is below BVDSS. Specifically, transistors 705 and N13 are in series, and transistors 704 and N14 are in series. However, depending on the BVDSS which is governed by the process technology, transistors 704 and 705 may not be needed.
The latch also has two PMOS transistors 714 and 715, both having a W/L of 39/10. These PMOS transistors connect the latch to ERSEL and ERSELb lines, which are complements of each other. By using these lines, data may be stored into the latch as desired. PMOS type devices are used so that negative voltages do not pass through. In particular, transistors 714 and 715 block voltages at ER1 and ER1B from passing through to ERSEL and ERSELb respectively. At ERSEL and ERSELb are inverter drivers, which in an embodiment is driven using a 3-volt supply. Voltages at ER1 and ER1b may be 12 volts or negative at different times of the operation. Transistors 714 and 715 prevent the voltages at ER1 and ER1 b from passing through to damage transistors 714 and 715.
Connected to ERSEL and ERSELb are logic gates governed by input signals ERSEbC, SSELb, and N10SWbH. Circuit block 722 is a voltage level shifter, which may change the signal level from one voltage level to another. The following table A gives the ERSEL and ERSELB outputs based these signals.
0 V (after
The upper supply pull-up circuit has two PMOS transistors, both having sources connected to an upper voltage supply line VCC12. The well connections for the devices are connected to VCC12. In an embodiment, VCC12 may be 12 volts or 10 volts, switching between the two as needed. VCC12 may be provided by an on-chip voltage pump circuit, but may also be provided from an source external to the integrated circuit or supply. Further in this specific embodiment, both PMOS transistors are the same size, having a W/L of 6/10. VCC12 is supplied by a current limited source circuit. The current limiter limits the current to about 10 microamps at about 12 volts. The current limited amount may be adjusted by changing the size of the devices in the current limiter circuit. For example, making the current lower will make the latch slower, but reduce current drain. The current limiter may limit the current to any desired value. For example, the current limiter may limit current to 100 microamps or less. The less current means the less power that will be consumed, but typically the performance decreases too.
The lower supply pull-down circuit has two NMOS transistors, both having sources connected to a lower voltage supply line VEE5. VEE5 may be −5 volts or −20 volts, switching between the two as needed. VEE5 may be provided by an on-chip voltage pump circuit or supply, but may also be provided from a source external to the integrated circuit.
The well connections for the devices are connected to a DNW1 line. In this specific embodiment, both NMOS transistors are the same size, having a W/L of 5.8/10. VEE5 is provided by a current limited supply, which limits the current to about 10 microamps at about −5 volts. The current limiter may limit the current to any desired value. For example, the current limiter may limit current to 100 microamps or less. The less current means the less power that will be consumed, but typically the performance decreases too.
In operation, the latch as described in
In an embodiment, the
In an embodiment, all logic gates are operated at a 3-volt supply. One mode is a decoding or set latch mode. When SSELb is 0 volts, this means this latch is selected. During this mode, N10SWbH is 3 volts and ERSELbC is 3 volts, so that ERSEL is 0 volts and ERSELb is 3 volts. In read/program operations, the signals N10SWbH and ERSEbC remain high (3 volts) so that ER is −5 volts and ERb is 12 volts. Then an associated memory cell M1, M2, M3, or M4 will pass p2pos (0:3) to the array to achieve the read or program operation. In erase operation, after the latch was set and the erase command was recognized by the state machine, ERSEbC is 0 volts. So the latch changes its setting, selected latch ER is 12 volts, and ERb is −5 volts so select CG<X>(p2<X.) will pass −20 volts to the array to erase operation perform when the state machine enables a −20-volt pump. VB2 changes its level from 3 volts to −10 volts. VCC12 changes its level from 12 volts to 0 volts. VEE5 changes its level from −5 volts to −20 volts. During a decode mode, program mode, or read mode, VCC12 is 12 volts and VEE5 is −5 volts. During an erase mode, VCC12 is 0 volts and VEE5 is −20 volts. Supply voltages for the latch are summarized in table B below.
Each array has n+1 rows, where n is an integer. There are n+1 word lines, WL0 to WLn. At an end of each array, a latch of the invention is connected to the array. A group of 64 rows of memory cells is called a segment. In this embodiment, each latch is laid out in the same pitch as one segment of memory cells. In this embodiment, a segment is defined as 64 rows of memory cells. However, in other embodiments, there can be any number of memory cell rows (or word lines) per segment, such as 1, 2, 8, 16, 32, 80, 128, and others. Since there are two arrays, there are two latches. In other embodiments, two latches may be placed into one block, adjacent to both arrays, instead of two blocks.
There are a number of latches connected in parallel. Each latch in column 807 is connected to an upper current-limited supply line 808 and lower current-limited supply line 810. Supply
There is one upper current limiter circuit 908 and one lower current limiter circuit 918 for the latches in the column. The latches share the upper current limiter and lower current limiter. The latches and current limiters are between pumps (or voltage sources) A and B. In other embodiments of the invention, there may be a single pump (or voltage source) connected to the current limiter circuits, instead of two pumps (or voltage sources). The latches L1, L2, L3, and L4 may be formed close or adjacent to the array of memory cells, while the current limiter circuity may be farther away from the array. This figure shows a separate pump or source of current limiter circuit. However, in other embodiments, these may be integrated or the same circuitry.
There may be any number of latches, not only four as shown in the figure. For example, there may be 2, 3, 4, 8, 16, 20, 23, 58, 62, 128, 256, or more latches. There may also be any number of memory cells. In a specific embodiment, there will be one latch for each segment of rows of memory cells. Programming may proceed line by line (e.g., row by row, column by column, sector or sector) by connecting the appropriate or corresponding latch to the control lines of a selected memory cell. Each of the memory cells may be Flash or EEPROM cells, or other type of nonvolatile cells, some of which were listed above.
Since one intended use of these latches is for decoding during programming or erase of Flash or other nonvolatile memory cells, they do not need to be very fast. But there are typically many latches like these on a single integrated circuit, all operating at the same time. It is important that each consumes relative low power. It is important that each latch uses a minimal amount of integrated circuit area. Also, with many latches switching at the same time, this may cause an amount of noise. By current limiting the latches, this reduces the amount of noise. If these current limited latches are used in other embodiments, they can be designed to achieve or obtain the desired benefits. The input of the latch can be any level or any polarity of the voltages.
A further benefit of the parallel configuration of current-limited latches of
In an embodiment, because of the range of voltages used (both positive and negative), the NMOS transistors used in the latch 702 of
In a specific embodiment, for the latch, the NMOS devices or transistors are deep n-well devices and the PMOS devices or transistors are triple n-well devices. In such a case, a DNW2 line provides a voltage to the deep n-well region of the devices.
To handle a partitioned array, each latch of the invention includes four pass gates or switches, shown in
In brief, the circuitry provides current-limited supplies. The magnitude of the current may be adjusted by changing the size of the transistors. For example, a transistor connected to VEE5_U has a W/L of 20/8. By altering this size, the amount of the current may be adjusted as desired. Similarly, VCC12_U and VCC12_D may be adjusted.
This description of the invention has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form described, and many modifications and variations are possible in light of the teaching above. The embodiments were chosen and described in order to best explain the principles of the invention and its practical applications. This description will enable others skilled in the art to best utilize and practice the invention in various embodiments and with various modifications as are suited to a particular use. The scope of the invention is defined by the following claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4342102||Jun 18, 1980||Jul 27, 1982||Signetics Corporation||Semiconductor memory array|
|US5095344||Jun 8, 1988||Mar 10, 1992||Eliyahou Harari||Highly compact eprom and flash eeprom devices|
|US5270979||Mar 15, 1991||Dec 14, 1993||Sundisk Corporation||Method for optimum erasing of EEPROM|
|US5278785||Jun 12, 1991||Jan 11, 1994||Emanuel Hazani||Non-volatile memory circuits and architecture|
|US5313432 *||Nov 12, 1991||May 17, 1994||Texas Instruments Incorporated||Segmented, multiple-decoder memory array and method for programming a memory array|
|US5334890||Oct 30, 1992||Aug 2, 1994||United Memories, Inc.||Sense amplifier clock driver|
|US5380672||Sep 3, 1993||Jan 10, 1995||Sundisk Corporation||Dense vertical programmable read only memory cell structures and processes for making them|
|US5602987||Dec 29, 1993||Feb 11, 1997||Sandisk Corporation||Flash EEprom system|
|US5668758||Jan 26, 1995||Sep 16, 1997||Macronix Int'l Co., Ltd.||Decoded wordline driver with positive and negative voltage modes|
|US5712180||Feb 28, 1996||Jan 27, 1998||Sundisk Corporation||EEPROM with split gate source side injection|
|US5745410||Nov 17, 1995||Apr 28, 1998||Macronix International Co., Ltd.||Method and system for soft programming algorithm|
|US5777924 *||Jun 5, 1997||Jul 7, 1998||Aplus Integrated Circuits, Inc.||Flash memory array and decoding architecture|
|US5991517||Dec 20, 1996||Nov 23, 1999||Sandisk Corporation||Flash EEprom system with cell by cell programming verification|
|US6005802||Oct 7, 1998||Dec 21, 1999||Kabushiki Kaisha Toshiba||Nonvolatile semiconductor memory device using a bit line potential raised by use of a coupling capacitor between bit lines|
|US6104665||Dec 4, 1998||Aug 15, 2000||Macronix International Co., Ltd.||Enhanced word line driver to reduce gate capacitance for low voltage applications|
|US6222762||Aug 7, 1997||Apr 24, 2001||Sandisk Corporation||Multi-state memory|
|US6230233||Sep 13, 1991||May 8, 2001||Sandisk Corporation||Wear leveling techniques for flash EEPROM systems|
|US6285578||Jan 5, 2000||Sep 4, 2001||Industrial Technology Research Institute||Hidden refresh pseudo SRAM and hidden refresh method|
|US6356481 *||Jun 16, 2000||Mar 12, 2002||Stmicroelectronics S.R.L.||Row decoder for a nonvolatile memory with capability of selectively biasing word lines to positive or negative voltages|
|US6380636 *||Jan 19, 2000||Apr 30, 2002||Mitsubishi Denki Kabushiki Kaisha||Nonvolatile semiconductor memory device having an array structure suitable to high-density integrationization|
|US6445621||Apr 11, 2000||Sep 3, 2002||Mosel Vitelic, Inc.||Dynamic data amplifier with built-in voltage level shifting|
|US6580629||Oct 2, 2001||Jun 17, 2003||Texas Instruments Incorporated||Semiconductor device array having dense memory cell array and hierarchical bit line scheme|
|US6639824 *||Sep 19, 2002||Oct 28, 2003||Infineon Technologies Aktiengesellschaft||Memory architecture|
|US6704241||Sep 6, 2002||Mar 9, 2004||Winbond Electronics Corporation||Memory architecture with vertical and horizontal row decoding|
|US6806525 *||Oct 24, 2002||Oct 19, 2004||Kabushiki Kaisha Toshiba||Semiconductor device and operation method thereof|
|US6818937 *||Jun 4, 2002||Nov 16, 2004||Micron Technology, Inc.||Memory cell having a vertical transistor with buried source/drain and dual gates|
|US6871258 *||May 30, 2002||Mar 22, 2005||Stmicroelectronics S.R.L.||Method for erasing an electrically erasable nonvolatile memory device, in particular an eeprom-flash memory device, and an electrically erasable nonvolatile memory device, in particular an eeprom-flash memory device|
|US7042752||Dec 3, 2003||May 9, 2006||Renesas Technology Corp.||IC card system using photo-detectors for protection|
|US20010050877||Jun 1, 2001||Dec 13, 2001||Matsushita Electric Industrial Co., Ltd.||Semiconductor memory|
|US20020008279||Dec 10, 1998||Jan 24, 2002||Tsukasa Ooishi||Semiconductor device|
|US20020015344||Sep 25, 2001||Feb 7, 2002||Monolithic System Technology, Inc.||High-speed read-write circuitry for semi-conductor memory devices|
|US20020031029 *||Oct 2, 2001||Mar 14, 2002||Yoshihiro Ogata||Semiconductor device array having dense memory cell array and hierarchical bit line scheme|
|US20020149048||Jun 4, 2002||Oct 17, 2002||Noble Wendell P.||Memory cell having a vertical transistor with buried source/drain and dual gates|
|US20030028709 *||May 30, 2002||Feb 6, 2003||Stmicroelectronics S.R.I.||Method for erasing an electrically erasable nonvolatile memory device, in particular an EEPROM-FLASH memory device, and an electrically erasable nonvolatile memory device, in particular an EEPROM-FLASH memory device|
|US20030067804 *||Aug 20, 2002||Apr 10, 2003||Giovanni Guaitini||EEPROM flash memory erasable line by line|
|US20040047224 *||Sep 6, 2002||Mar 11, 2004||Winbond Electronics Corporation||Memory architecture with vertical and horizontal row decoding|
|US20040070021||Oct 9, 2002||Apr 15, 2004||Yuan Jack H.||Flash memory array with increased coupling between floating and control gates|
|US20050101236||Dec 20, 2004||May 12, 2005||Sandisk Corporation||Current-limited latch|
|USRE37419 *||Oct 29, 1999||Oct 23, 2001||Aplus Flash Technology Inc.||Flash memory array and decoding architecture|
|EP1152538A1||Jan 6, 1999||Nov 7, 2001||Fujitsu Microelectronics Europe GmbH||Mixed-signal circuitry and integrated circuit devices|
|1||"Communication Relating to the Results of the Partial International Search", corresponding PCT application No. PCT/US03/38328, International Searching Authority, European Patent Office, Jan. 6, 2004, 5 pages.|
|2||"Notification of Transmittal of the International Search Report or the Declaration", corresponding PCT application No. PCT/US03/38328, International Searching Authority, European Patent Office, Jan. 18, 2005, 10 pages.|
|3||European Patent Office, Substantive Examination Report mailed in corresponding European Patent Application No. 03 812 793.2 on Jul. 2, 2007, 3 pages.|
|4||International Search Report, PCT/US03/38328 filed Jan. 12, 2003.|
|5||Office Action dated May 1, 2006 in corresponding U.S. Appl. No. 11/019,990, 10 pages.|
|6||USPTO, "Notice of Allowance", mailed in related U.S. Appl. No. 11/019,990 on Oct. 24, 2005, 4 pages.|
|7||USPTO, "Notice of Allowance", mailed in related U.S. Appl. No. 11/019,990 on Sep. 18, 2006, 4 pages.|
|8||USPTO, "Office Action", mailed in related U.S. Appl. No. 11/019,990 on Apr. 27, 2007, 9 pages.|
|9||USPTO, "Office Action", mailed in related U.S. Appl. No. 11/019,990 on Jun. 9, 2005, 8 pages.|
|10||USPTO, Notice of Allowance mailed in continuation U.S. Appl. No. 11/019,990 on Sep. 4, 2007, 5 pages.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8248881 *||Aug 21, 2012||Micron Technology, Inc.||Method and apparatus for managing behavior of memory devices|
|US8432765||Apr 30, 2013||Micron Technology, Inc.||Method and apparatus for managing behavior of memory devices|
|US20100195430 *||Aug 5, 2010||Roohparvar Fariborz F||Method and apparatus for managing behavior of memory devices|
|US20120314507 *||Dec 13, 2012||Micron Technology, Inc.||Reduced voltage nonvolatile flash memory|
|U.S. Classification||365/185.11, 365/226, 365/227|
|International Classification||G11C16/12, G11C16/04|
|Dec 6, 2002||AS||Assignment|
Owner name: SANDISK CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:WANG, CHI-MING;CHEN, KUO-LUNG;TSAO, SHOUCHANG;REEL/FRAME:013588/0451
Effective date: 20021204
|May 20, 2011||AS||Assignment|
Owner name: SANDISK TECHNOLOGIES INC., TEXAS
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:SANDISK CORPORATION;REEL/FRAME:026317/0397
Effective date: 20110404
|Aug 3, 2011||FPAY||Fee payment|
Year of fee payment: 4
|Aug 19, 2015||FPAY||Fee payment|
Year of fee payment: 8